Differential effects of commensal bacteria on progenitor cell adhesion, division symmetry and tumorigenesis in the Drosophila intestine

Microbial factors influence homeostatic and oncogenic growth in the intestinal epithelium. However, we know little about immediate effects of commensal bacteria on stem cell division programs. In this study, we examined effects of commensal Lactobacillus species on homeostatic, and tumorigenic stem cell growth in the Drosophila intestine. We identified Lactobacillus brevis as a potent stimulator of stem cell growth. In a wildtype midgut, Lactobacillus brevis activates growth regulatory pathways that drive stem cell divisions. In a Notch-deficient background, Lactobacillus brevis-mediated growth causes rapid expansion of mutant progenitors, leading to accumulation of large, multi-layered tumors throughout the midgut. Mechanistically, we showed that Lactobacillus brevis disrupts expression and subcellular distribution of progenitor cell integrins, supporting symmetric divisions that expand intestinal stem cell populations. Collectively, our data emphasize the impact of commensal microbes on growth and maintenance of the intestinal progenitor compartment.


34
Drosophila melanogaster is a popular system to study microbial control of ISC growth due to an extensive 35 toolkit for host genetic manipulation, and a simple, cultivable microbiome that is easy to modify (Broderick and 36 Lemaitre, 2012;Koyle et al., 2016). Importantly, key regulators of ISC growth are evolutionarily conserved

73
To study bacterial effects on intestinal tumors, we used the temperature-controlled escargot-GAL4, GAL80 ts , 74 UAS-GFP (esg ts ) transgenic fly line to express an inducible Notch RNAi construct (UAS-N RNAi ) in ISCs and 75 enteroblasts (collectively referred to as progenitor cells) at the restrictive temperature of 29C. Intestines of 76 control esg ts /+ females contained evenly distributed GFP-positive progenitors and prospero-positive 77 enteroendocrine cells in a simple epithelium dominated by large, polyploid enterocytes (Fig. 1B, Fig. S1A). In line 78 with an earlier study that described massive stem cell growth in flies with Notch-deficient progenitors (Patel et

83
In contrast to intestines with a conventional microbiome, Notch-deficient tumors rarely appeared in age-84 matched, germ-free (GF) flies, indicating microbial requirements for tumor growth (Fig. 1D), although we cannot 85 exclude the possibility that tumors eventually form in GF flies with age. To identify bacterial species that promote 86 tumors, we examined posterior midguts of adult esg ts /N RNAi flies that we associated exclusively with common 87 species of Lactobacillus commensals, a dominant genus within the fly microbiome (Adair et al., 2018;Wong et al., 88 2013). To focus exclusively on adult tumors, we raised esg ts /N RNAi larvae with a conventional microbiome under 89 conditions that prevent Notch inactivation. Upon eclosion, we fed adults an antibiotic cocktail that depleted the 90 bacterial microbiome below detectable levels, and re-associated flies with Lactobacillus brevis (Lb), or 91 Lactobacillus plantarum (Lp) (Fig. 1A). We compared each mono-association to conventionally reared (CR) 92 esg ts /N RNAi flies that contained a poly-microbial gut microbiota. Mono-association of esg ts /N RNAi flies with Lp 93 resulted in few visible tumors (Fig. 1E). In contrast, mono-association with Lb caused multiple, large tumors 94 throughout the posterior midgut (Fig. 1F), indicating that Lb is sufficient for tumor development.

6
We then quantified impacts of bacterial association on midgut tumors. First, we developed a four-point 96 system to classify intestines, ranging from no visible defects (level I) to intestines with progenitor and 97 enteroendocrine-rich tumors (level IV, Fig. S1D). In a blinded assay, we categorized 85% of CR esg ts /N RNAi intestines 98 as level IV, whereas only 20% of GF intestines belonged to the same category (Fig. 1G), confirming bacterial effects 99 on gut tumors. Consistent with our initial observations, GF and Lp-associated intestines had similarly mild levels 100 of midgut tumors (Fig. 1G). In contrast, all intestines associated with Lb had level IV tumors within five days of 101 Notch inactivation. To measure total tumor size per midgut, we quantified the posterior midgut area occupied by  Lb-associated flies are a result of increased tumor initiation, or accelerated tumor growth we quantified numbers 106 of tumors in each intestine. Similar to our assessment of tumor size, association with Lb had a significant impact 107 on tumor numbers, resulting in approximately three times as many tumors per gut as CR counterparts (Fig. 1I).

108
Collectively, our data indicate that association with Lb increases the frequency of midgut tumor initiation.

122
To determine which factors from Lb promote tumors, we measured tumors in flies that we continuously fed  (Fig. 2D). Finally, we noticed that esg ts /N RNAi flies mono-associated with Lb died significantly faster 128 than CR counterparts, while Lb did not shorten the lifespan of esg ts /+ controls (Fig. S2), arguing that cell wall 129 components from L. brevis promote initiation of Notch-deficient tumors, resulting in premature host death. To determine how Lb affects Notch-deficient progenitors, we used RNA-sequencing to identify the 140 transcriptional profiles of FACS-purified, GFP-positive progenitors from Lb-associated esg ts /+ and esg ts /N RNAi 141 intestines (Fig. 3A). As controls, we sequenced transcriptomes of esg ts /+ and esg ts /N RNAi progenitors from GF flies, 142 or flies that we mono-associated with Lp. Principal Component Analysis (PCA) revealed that Notch-deficient 143 progenitors segregate from wildtype progenitors along PC1, regardless of bacterial association (Fig. 3B).

144
Differential gene expression analysis showed that the majority of gene expression profiles altered by Notch-145 depletion were shared between GF, Lb-associated and Lp-associated intestines (

180
As bacterial association is higher in Notch-deficient intestines than wildtype intestines (Fig. 4A), and Notch 181 inactivation diminishes expression of IMD pathway components, we asked if IMD affects host association with Lb.

182
Consistent with a role for IMD in the control of intestinal Lb, we found that imd mutants had significantly higher 183 Lb loads than wild-type controls ten days after mono-association with Lb ( Fig. 4B). Conversely, constitutive 184 activation of IMD in enterocytes (Myo1A ts /ImdCA) reduced Lb load to approximately 4% of that found in imd 185 mutants ( Figure 4B). These data support a role for IMD in regulation of intestinal Lb. However, it is important to 186 note that the increased bacterial abundance in imd mutants is considerably less pronounced than increases 187 observed upon Notch depletion (compare Fig. 4A and 4B). Thus, we believe that additional, IMD-independent 188 mechanisms control bacterial numbers in esg ts /N RNAi intestines that require identification.

189
Finally, we measured the effects of Notch inactivation on host association with Lactobacillus commensals.

190
Here, we completed a longitudinal measurement of bacterial load in intestines of esg ts /+ and esg ts /N RNAi flies that 12 we mono-associated with Lp or Lb. In general, our data match earlier reports that total numbers of intestinal 192 bacteria increase in flies with age (Clark et al., 2015;Guo et al., 2014). In wild-type esg ts /+ intestines, the rates of 193 increase in host-association with Lp and Lb are nearly indistinguishable, with Lb associating to lower levels at all 194 times tested (Fig. 4C). Initially, Lb also associated with esg ts /N RNAi intestines to lower levels than Lp. However, we 195 noted substantial effects of Notch inactivation on subsequent progressions in host-microbe association. In this 196 case, association with Lp increased at a considerably slower rate than association with Lb ( Fig. 4D). Exponential

212
As Lb grows effectively in Notch-deficient intestines, where it promotes tumors, we reasoned that Lb will have 213 distinct growth-enhancing effects on progenitor cells. To test this hypothesis, we looked for progenitor cell 214 transcriptional events that were specific to association with Lb. Principle component analysis identified a 215 transcriptional response that is unique to Lb in wild-type and Notch-deficient progenitors (Fig. 2B, Fig. S3A,B).

216
Regardless of host genotype, association with Lb specifically increased expression of genes required for cell 217 growth, such as DNA replication, and mitotic spindle organization (

225
Given the positive effects of Lb on expression of growth regulators in esg ts /+ and esg ts /N RNAi progenitors, we 226 asked if Lb activates ISC division in wild-type progenitors. To answer this question, we mono-associated GF wild-14 type (esg ts /+) flies with Lb and quantified phospho-histone 3-positive (PH3+) mitotic cells in adult midguts. Similar 228 to effects on tumors, Lb stimulated growth of wildtype progenitors to significantly higher levels than CR, GF, or 229 Lp-mono-associated flies (Fig. 5C). Thus, our data indicate that association with Lb diminishes expression of genes 230 required for progenitor adhesion to the extracellular matrix, and induces expression of genes required for 231 epithelial growth, promoting ISC division in wildtype and Notch-deficient progenitors.

252
GF esg ts /+ flies, we detected basolateral enrichment of mys in GFP-positive progenitors (Fig. 6E). In contrast, and 253 similar to our electron microscopy results, we found that Lb colonization caused progenitors to round up and 254 adopt a more apical position within the epithelium (Fig. 6F). Furthermore, association with Lb had visible impacts 255 on mys localization, characterized by discontinuous basolateral distribution, and atypical apical enrichment of mys 256 (Fig. 6F, arrowheads). To directly measure effects of Lb on subcellular distribution of integrins, we developed an 257 immunofluorescence-based assay that allowed us to quantify apical:basolateral ratios of mys in progenitors (Fig.  17   S5). With this assay, we detected basal enrichment of mys in GF progenitors (Fig. 6H). Association with Lb shifted 259 the distribution of mys, resulting in significant increases in apical mys (Fig. 6H). To determine if Lb-dependent 260 effects on integrin subcellular distribution are downstream consequences of stem cell division, we blocked growth 261 in progenitors of Lb-associated flies by expressing the cell cycle inhibitor dacapo (esg ts /dap) (Fig. 6I). Notably, 262 when we examined growth-impaired midguts, we found that Lb continued to cause increases in apical mys (  (Fig. 7H). Combined with quantification of Dl+ stem cells (Fig. 7A-C), our data indicate that relative to CR or GF 295 flies, Lb shifts composition of the progenitor cell compartment towards stem cells.

296
We do not see elevated levels of cell death (Fig. S6A), or increased expression of apoptosis regulators (Fig.   297 S6B) within progenitors of Lb mono-associated flies compared to CR or GF controls. Thus, we do not believe that 298 Lb affects cell composition within the progenitor compartment by preferentially promoting enteroblast death 299 (Reiff et al., 2019). As an alternative, we tested the hypothesis that Lb increases stem cell numbers by promoting   344 Reedy et al., 2019). Likewise, Lb-derived uracil activates epithelial generation of ROS in adults, a noxious agent 345 that promotes tissue damage and repair (Lee et al., 2013). In addition, cell walls of Lp contain DAP-type 346 peptidoglycan that activates the IMD pathway with substantial effects on transcription (Broderick et al., 2014; 347 Lesperance and Broderick, 2020). In the future, it will be of interest to test the relationships between microbe-348 specific immune responses, ROS production, and epithelial growth. As poly-bacterial communities have distinct    (Patel et al., 2015), it would be pertinent to understand whether the actions of Lb on integrins is specific to 390 progenitors or if enterocytes are also affected.

391
How Lb disrupts integrins and promotes stem cell growth requires clarification. As stem cells derive cues from 392 the surrounding epithelium to direct their growth, we consider it likely that mature epithelial cells, such as

426
(100g/mL)) for 5 days at 25C. Conventionally reared controls were fed autoclaved food without antibiotics for 427 5 days at 25C. To generate monoassociated animals, flies were made germ free as above then were fed 1mL 428 OD600=50 of L. brevis or L. plantarum resuspended in sterile 5% sucrose/PBS on a cotton plug overnight at 25C.

429
During this overnight feeding, CR and GF controls were fed sterile 5% sucrose/PBS without bacteria. The following  were then shifted back to 29C for 8 days before dissecting and counting clones from 30 intestines per treatment.

441
The imd and UAS-ImdCA lines had been backcrossed into our wild-type w 1118 background. Drosophila lab stocks and have been previously described (Petkau et al., 2016).

478
To determine the apical and basal mys intensity, we drew a line of 10 pixel width from the basal side to the apical 479 (lumen side) side across GFP+ progenitor cells. We defined apical and basal progenitor cell borders as 50% of the 480 maximum GFP intensity, as this GFP intensity coincides with the basal mys intensity peak. We determined the 481 intensity of GFP and mys across the progenitors using the function plot profiles, copied these values into Excel

492
Samples were stored at -80°C until all samples were collected. RNA was isolated via a standard Trizol chloroform extraction and the RNA was sent on dry ice to the Lunenfeld-

502
Adaptors and reads of less than 36 base pairs in length were trimmed from the raw reads using Trimmomatic . In EdgeR, genes with counts less than 1 count per million were filtered and libraries were normalized 508 for size. Normalized libraries were used to call genes that were differentially expressed among treatments. Genes 509 with P-value < 0.01 and FDR < 5% were defined as differentially expressed genes.